REPORTS
A
Remarkable
Auroral
Event
on
Jupiter
Observed in the Ultraviolet with the
Hubble Space Telescope
J. C.
Gerard,*
D.
Grodent,
R.
Prange,
J. H.
Waite,
G. R.
Gladstone,
V.
Dols,
F.
Paresce,
A.
Storrs,
L.
Ben
Jaffel,
K. A. Franke
Two sets of ultraviolet images of the Jovian northaurora wereobtained with the Faint Object Cameraonboard the HubbleSpace Telescope. The first series showsanintense
discretearc in nearcorotation with the planet. The maximum apparent molecular hy-drogen emission rate corresponds toan electron precipitation of -1 watt per square
meter,which is about 30,000 times larger than the solar heating byextreme ultraviolet radiation. Suchaparticle heatingrateof the auroralupperatmosphere of Jupiter should cause alarge transienttemperatureincrease andgenerate strongthermospheric winds. Twenty hours after initial observation, the discrete arc had decreased in brightness by morethanoneorderof magnitude. The time scale and magnitude of the change in the ultraviolet aurora leads us to suggest that the discrete Jovian auroral precipitation is relatedtolarge-scale variations in thecurrent system,asis thecasefor Earth's discrete aurorae.
The first observations of the ultraviolet (UV) aurora on Jupiter were obtained by theUV spectrometer (UVS) onboard the Voyager 1 and 2 spacecraft in 1979 (1-3). Dataaccumulated since 1980 with the UV spectrographon the International Ultravi-oletExplorer (IUE) satellite (4) havebeen usedtoindirectly characterize themain fea-tures of the morphological and brightness distribution of the aurora. Fromthese early
observations, someinformation was derived
on its temporal behavior. Livengood et al.
(4) analyzed 10yearsof IUEdata and found
that the average observed auroralbrightness profile was generally stable withinafactor ofabout 2 to 3. The lack ofspatial resolu-tion of these nonimaging instruments did not allow the determination of whether morphological changes (in auroral shapeor latitude) were associated with the bright-ness variations (4, 5). Some morphological differences were observed in 1992 on two Hubble SpaceTelescope (HST) Faint Ob-jectCamera (FOC) images (6) withnearly identical central meridian longitudes sepa-ratedby about3days. However,inthis case, both themaximumlocalbrightnessand the integrated radiated power showed little variation. The question of temporal
vari-ability is of major importance in
under-J. C.G6rard,D.Grodent,V.Dols,Laboratoire de Phy-sique Atmosph6rique et Plan6taire, Institut d'Astro-physique, Universit6 de Libge, Belgium.
R.Prang6,Institutd'AstrophysiqueSpatiale and Institut
d'AstrophysiquedeParis,Universit6dePans Sud, Orsay, France.
J. H.Waite, G. R.Gladstone,K. A.Franke,Southwest Research Institute, San Antonio, TX, 78228-0570, USA. F.Paresceand A.Storrs, Space Telescope Science In-stitute, Baltimore,MD21218, USA.
L.B.Jaffel,Institutd'AstrophysiquedeParis, France. *Towhomcorrespondence shouldbe addressed.
standing the origin and the acceleration mechanisms of theauroralparticlesexciting the JovianUV aurora.
Basicdifferences exist between the mag-netospheres of Jupiter and the Earth. The Earth's magnetosphere dynamics, con-trolledby the solar wind dynamo, organize auroral processes in a local time frame of reference with peak activity in the
mid-nightsector. These processes take placeon
magnetic field lines from the central to distant plasma sheetin the nightside mag-netotail. By contrast,the muchmore exten-siveJovianmagnetosphere isin quasi-coro-tation with the planet up to distances of about20Jovian radii. This featureexplains why theUVSexperimentonboard Voyager (3) didnotdetect any significantday-night variation.
In July 1993, two series of three HST FOC images, each oftheUV aurora, were taken nearly 20 hours apart to investigate the question of temporal variability. The filter isolateda20-nm-wide region centered on153 nm (7), whichisdominatedbythe emissionofthe molecularhydrogenLyman bands(H2 Lyman) and thecontinuum.The relevant parameters ofthe observations are specified inTable 1.
Unexpectedly, the first series of expo-sures recorded a very bright auroral event that gave rise to a FOC count level about four times higher than any previous obser-vations made with the same instrumental configuration. Nearlyparallel spectrograph-ic measurements made with the IUE (8) reveal that the emitted UV auroral radia-tion reached the second highest level re-corded in over 12 years ofIUEJovian
au-roral observations. In the first exposure of the first series (I1) (Fig. 1A), abright but 1994
longitudinally limitedportionof the auroral arc isvisible nearthe approaching (dawn) limb of the planet. In the third exposure (13) (Fig. 1B), the bright arc is considerably moreextended in longitude.A comparison of the location of theemission morphology on the two exposures (9) shows that the entire pattern is in quasi-corotation with the planet. However, the leading edge of the bright arc is shifted to smaller longi-tudesby -520, whichisless than the 650of planetary rotation. This slippage may be caused by a temporal brightness change along the auroral oval over the 107 min separatingthetwo images.
The quasi-corotation demonstrates the predominantly longitudinal control of the auroralemissionofJupiteralready suggested by the first FOC images at the Lyman ot frequency, obtained in 1992, (10) and by years of IUE observations (4). The aurora has an uncertain brightness distribution along the auroral oval and may have a local
time dependence as well as a longitude
dependence. We also expect these H2
Ly-man emissions to be considerably brighter nearthe limb (1 1).Forexample, the appar-entbrightnessnearthe limb isexpectedto be abouteighttimesbrighter thanan iden-ticalaurorawould be on thecentral merid-ian at alatitude of60%N.
Inthebackground-subtracted versions of images12andW1 (Fig. 2,Aand B, respec-tively), the bright arc seen in Fig. 1A stretches from a latitude of 60%N and a systemIIIlongitude
(X111)
of 1800to-700N onthe dawn limb.Asecond,weaker spot of emission extends along the centralmeridi-an. Weak, diffuse emission fills the entire
polar cap and reaches -50%N near = 190°. The brightest pixels correspond to a count rate of -13 counts per pixel. This level is higher than previous HST FOC observations (6, 12, 13), which typically reached 2 to 4 countsper pixel above the
diskbackgroundforsimilar exposure times.
Converted into an apparent emission rate, themost intenseparts ofthe auroralarcs in imagesI1, I2, and 13 reach -3.6
megaray-leighs (1 MR = 1012 photons per square
centimeter per second) of H2 Lyman
Table 1. Main characteristics of the HST FOC auroral observations. Abbreviations:UT,universal time;X,wavelength;CML,central meridian longi-tude;NUV,nearultraviolet.
Date UT Central
CMVL
ImgX (nm) (degrees) Image 17July 00:47 210 51 NUV 02:21 153 108 11 03:48 153 160 12 04:08 153 172 13 23:16 153 146 W1 18July 00:41 153 198 W2 01:01 153 210 W3 m on October 27, 2018 http://science.sciencemag.org/ Downloaded from
emission, equivalentto -6 MRoftotal H2 emission (14).
Figure 2Billustrates the dramatic change that occurred during about twoJovian ro-tations. The bright arc previously seen at
Kin1
> 180° faded toa much weaker zoneof inhomogeneous emission with a maximum near 600N, KX = 2000. Table 2 shows a comparison of the total power radiated in H2 Lyman in images 12 and W1 and the maximum count rates in both exposures. The peaks of the auroral emission are in a ratioofnearly 8 to 1. However, the bright-estpart ofthe intense arc in Fig. 2A maps into a weak zone of -1 count per pixel, indicating a drop to 1/13 of the local emis-sionrate.Aglobal view of the auroral morphology and its variation during the event is best givenbycomposite polar orthographic pro-jections of images I1, I2, and 13 (Fig. 2C) and W1, W2, andW3 (Fig.2D); they illus-tratehow theoverall decrease in emission affected the morphology of the emission.
The bright aurora observed in the first
group of images (Fig. 2C) lies close to but not along the oval described by the foot-print of the magnetic field lines crossing
the equatorial plane at 30 Jovian radii
(Rj).
Gerardet al. (12) mapped the aurora observed over a complete Jovian rotation andfound thatabetter fit canbeobtained by shifting the center of the model ovalby a few degrees. The leading edge of the bright emission lies near KX = 1800, al-thoughaweaker extension isobserved up to 1650. Asecondary weaker alignmentisseen along the 160° meridian, extending from 62%N up to the limit ofvisibility. Dimmerunstructured aurora is visible inside and
outside the discreteemission. Itreachesan
equatorward limitof -50%N at
K,,,
= 180°atthe lr threshold applied tothe individ-ual images. The corresponding projection 20 hours later (Fig. 2D) shows, in agree-mentwith Fig. 2B, considerable brightness variations, but the morphology is reminis-cent of Fig. 2C. Aweak emission band is still presentparalleltoandoutside of the30
Ri
ovalatK111
> 180°.These featuresleadto the conclusion that the bright auroral arcalready observed previously in the sector
K111
> 180°overlaidadiffuse auroral region that remained nearly invariant in bright-nessthroughout theevent.Table 2liststheestimated totalradiated power over the halfhemisphere facing the Earthderivedfrom both 12 andW1. Using an energy conversion of -7 for electron deposition (assuming that the auroral par-ticlesareenergeticelectrons)in aH2 atmo-sphere (15-17), wefind that the 6 MRof total H2emissioncorrespondto alocalflux of -1 Wm- 2 and atotal radiatedpowerof 1012 W. The input ofsuch a high energy flux must considerably perturb the energy
1676
Fig. 1. Images of the Jovian auroraobtained at awavelength of 153nm with the HST FOC. The brightness of the diskbackground from solar UV radiation scattered by the atmosphere decreases from themid-latitudes (lowerleft)tothepolar region. The planetary limb is visible against dark space. (A) The firstimageof the first sequence (exposure11). The aurora is visible near the limb and forms an ansa or loop near the(astronomical) eastern edge of the oval. (B) The third image (exposure 13) shows how the auroral distribution evolved in the 107minseparating thetwoexposures.
balance andthe temperature profile ofthe Jovian upperatmosphere.Asenergetic elec-trons interact with the hydrogen gas by elastic andinelastic collisions, a large frac-tion ofthe energy input is converted into local gas heating. The consequences ofan electron power input of-1 Wm-2 canbe
crudely estimated with simple
one-dimen-sional (1D) energydegradation and energy balancemodels(16, 18).
Because no direct information on the energydistribution of the auroralparticlesis available,webaseourestimatesonthe HST spectroscopy observations of Trafton et al. (19) with a corresponding spectral power index of -4 andalow-energy cutoffat 22 keV.Two situations maybeanalyzed: (i) A 1Dnumericalmodelprovidesanupper lim-itofthe temperature, and.(ii) a more real-istic concept that includes the time re-sponse and horizontal transport of the 3D
atmospheric system is considered. The
high-altitude (exospheric) temperature for
the case of 1 W m- 2 calculated at steady statewitha 1DmodelbyWaiteetal. (16) is 230,000 K, which is much higher than the observed H3+ temperature of 1000 to
1500K(20).Themolecular diffusiontimes
just above the methane homopause where the maximum heating occurs are much longer thanaJovian day.Therefore,wecan examine a zonally averaged heating rate, which leadsto areductioninthevertically
integratedheat flux(15 ergscm-2
s-1)
anda corresponding exospheric temperature of
3000 K.Thevertical temperatureprofilefor thiszonallyaveragedcaseisshowninFig.3 along with an observationally constrained profilefor a moretypicalaurora(19) (10to 100less thantheextremely brightauroraof Figs. 1 and 2). The time-variable
tempera-ture profile is bounded by these two
ex-SCIENCE * VOL.266 * 9DECEMBER 1994
Table 2.Local andintegratedemission and
en-ergyflux.
12 W1
Maximum localcount 13.3 1.7
(counts perpixel)
Maximum apparent >3.6 0.46 emissionrateLyman
bands and continuum (MR)
Local energy flux(Wm-2) >0.75 0.1
Total counts 18,100 5,400
Observed radiated power 1 0.3 (H2bands)(1012W)
Estimatedprecipitated 7 2
power(electrons)
(1012W)
treme cases. The initial temperatures may have risen locally to over 100,000 K. One can speculate that transport subsequently redistributed the heat by means of advec-tive processes. Finally, a high exospheric temperature of over 3000 K quite likely
persisted for several hours at all auroral
longitudeswith infrared hot spot structures
remaining over the initial zone of heating
andsignificant winds generated as aresult
of the large energy input.
An additional consideration is the un-certainty inourknowledgeof the spectrum oftheincomingprecipitatingparticles. The exospheric temperature result isquite sen-sitive tothelow-energy portion of this spec-trumbecause theprecipitatingparticles de-posittheir energyathigh altitudes(pressure levels of about 1 nbar) that are further removedfrom thehydrocarboncooling lay-er. For example, for 215-eV electrons, en-ergyinfluxes onthe order of100ergscm-2 s-1 produce exospheric temperatures of over
106
K. At these temperatures, the at-'I m 1111 11111 111111lood'.lIn--11 lm 11 111111m il 1111 1111111,110
on October 27, 2018
http://science.sciencemag.org/
10-5
a-105
.in101
Fig.2. (Aand B)Background-subtracted version oftheauroral imageswith alatitude-longitude grid
coordinate (in red). Only features >1o above thebackgroundarecolor-coded. Latitude parallelsare
separated by 100 and meridians increaseby 20°from the lowerrighttotheupperleft. Thewhitedots
indicate the footprintofthe30-Rjfield linescalculatedfromtheGoddard SpaceFlightCenter(GSFC) °6 magnetic field model(27). (A) Exposure12showinganexceptionally brightaurorawithacomplex intensity
distribution.(B)Exposure W1 showingadramaticchangeinauroralbrightnessoverthe19.5hourssince
theobservation in(A). (CandD)Polarprojectionsof thetwosetsofbrightand weakauroralimageswhere
datapoints >1aabove the diskbackgroundareconsidered.SystemIlllongitudesincrease in 200steps
clockwise from the reference meridian(X11= 0°),which is orientedtoward thetop. The 1800 meridian is
towardthe bottom. (C)Acompositeof thebright images11, 12,and13,and (D)acombination of the
weaker imagesW1,W2, and W3.
mosphere is no longer inhydrostatic
equi-librium and literally blows off, creating a
hydrogen corona.A very high temperature
mayhelp explainthehighly Doppler-broad-ened Lyman atprofiles that have been ob-served by IUE and more recently by HST
(21). Theymayalsoexplain whytheaurora
seems to often occur near the homopause
level rather than in the thermosphere (where cooling byhydrocarbons is negligi-ble), particularly at its brightest intensity
(22). Indeed, the only stable situation in
thecaseofenergy inputover50 ergscm-2
S-' iswhenthe altitude of the heating and coolingratesnearly coincide.
In the case of the previous exceptional
Jovian auroral eventobserved with IUE, it
wasdemonstrated that the auroral
enhance-mentcoincidedwiththe arrivalatJupiter of
a solar density disturbance, identified as a
coronal mass ejection (5). Therefore, we
searched for similar associations with the July 1993event.TheUlyssesspacecraftwas
at high heliographic latitude in July 1993, and theplasmadataconfirm thatitwasout of the streamer belt at the time of our
observations (23). Therefore, the Ulysses dataare notusefulindetermining the solar windconditionsatJupiter duringtheevent. The characteristic time of the observed brightnessvariationwas less than20hours.
Thisfactor, togetherwith the auroral
loca-tion at high latitudes (30
Rj),
indicates that the origin of auroral particles is not directlyconnectedtothe loplasmatorus (6 R.). Rather, they appear to originate fromthe more distant (middle) magnetosphere
andmaybe linkedtofield-alignedcurrents, observedduringtheUlyssesencounterwith
the Jovian magnetosphere, associated with
102 103
Temperature (K)
Fig. 3. Steady-state vertical distribution of the
temperatureof the Jovian auroral atmosphere
cal-culated forelectron precipitation with a 22-keV
cutoffandforatotal integrated heatingrateof15
ergs cm-2S-1 (solid line). For comparison, the profile derived from observational spectroscopic
constraints(20) is also shown(dotted line).
a high-latitude auroral arc observed at the Lymanat wavelengthwiththeHST(24). It
is interesting to compare the auroral pro-cessesatJupiter with the better document-ed and understood terrestrial counterpart. The above arguments (high latitude, short
time scale) suggest field-aligned current-driven auroral precipitations analogous to theterrestrial discrete aurora.
The strongly enhanced auroral emission
reportedherewasobservedondaysidefield
lines andwas essentiallyfixed inmagnetic
longitude,notinlocaltime.Longitudinally fixed auroralforms are consistentwith the
dominance of the corotational convective
flow within the Jovian magnetosphere. In
the inner magnetosphere (<20 R.), the plasmamotion is dominatedby the
corota-tion electric field generated by the rapid
rotation of the Jovian atmosphere-iono-sphere. Outside of this distance (20 R.), theplasmaaccelerationtimebecomes
long-er than theplasmaoutflow time. This
sug-gests a decoupling of the ionosphere from
the magnetosphere, that is, a departure
from rigid corotation (25). This presents the possibility that reconnection processes
nearthe magnetopause mayproduce local-izedsolar windconvectioncellsintheouter magnetosphere, which may resultin shears
in theplasmaflow neartheplasma
corota-tionboundary, inturn producing the field-alignedcurrentsthatareresponsibleforthe
high-latitude aurora. Evidence therefore
suggests that auroralprecipitations, similar-lydrivenonJupiter and the Earthby
field-aligned currents, nevertheless originate
fromdifferent mechanisms.
REFERENCESAND NOTES
1. A.L. Broadfoot etal.,J. Geophys. Res.86, 8259
(1981).
2. B. R.Sandel etal., Science206,962(1979). 3. F.Herbert, B. R.Sandel,A.L. Broadfoot,J.
Geo-phys.Res.92,3141(1987).
4. T. A. Livengood, H. W. Moos,G. E. Ballester, R.
'I~ ~ ~ I A,/ Mu I u - -a 104 on October 27, 2018 http://science.sciencemag.org/ Downloaded from
Prang6,Icarus97, 26(1992).
5. R.Prang6et al., J. Geophys.Res.98, 18779(1993). 6. J.C.G6rard, V. Dols, F. Paresce, R. Prang6, ibid., p.
18793.
7. The overall sensitivity of the optical system is 4.1 x 10-6counts per pixel per second for every kiloray-leigh of H2 Lyman emission. These observations were made before the addition of the Corrective Op-tics Space Telescope Axial Replacement (COSTAR), so the aberration of the HST primary mirror spread the light ofa point source over a circular halo 2 arc secin radius. All exposures (896s each) described here were made with the use of the same guide star for telescope pointing and tracking. Thecenter of the camera field of view was located along theplanet's central meridian at 600N during each image. 8. F. W. Harris, personal communication.
9. The precise location of the planetary limb in theFOC field of view was determined from both the near-UV exposure (Table 1), which shows no auroral emis-sion, and from the well-defined limb on each153-nm auroral image.
10. V. Dols, J. C. G6rard, F. Paresce, R. Prang6, A. Vidal-Madjar, Geophys. Res. Lett. 19,1803 (1992). 11. The bulk of the H2 Lyman emissions are emitted in
bands near 160 and 161nm and are optically thin in H2. Because thenorthern auroral oval is offset from the rotational pole, the effect of the slant path bright-ening is reduced somewhat by smearing during the exposure. During a typical exposure of 15min,the planet rotates by about 9° of longitude. The slant path enhancement is also reduced if the actual au-roral structure is unresolved by the FOC. A resolution of about 0.1 arc sec implies that structures smaller than about 400 km are diluted. For comparison, the scale height near the homopause on Jupiter is likely to be about 100 km or so, and a degree of latitude at 600N subtends about 600 km. We assumed that the aurora's vertical emission rate can be represented by a Chapman profile, used the nominal value of 100 km for the atmospheric scale height and an auroral height of 500 km above the cloud tops, and account-ed for the dilution causaccount-ed by the spatial resolution of the FOC (but not for smearing from therotation of Jupiter).
12. J. C.G6rard,V. Dols, R.Prang6,F.Paresce, Planet. SpaceSci., in press.
13. J. Caldwell, B. Turgeon, X.-M. Hua,Science 257, 1512 (1992).
14. This value can be considered as a lower limit be-cause the aberrated point spread function of the telescope and the blurring effect of the planetary rotation spreads the light of a point source over many pixels and decreases the apparentemissionrateof localized emission. Arough estimate of this effect may beobtained from thedeconvolved (Wiener) ver-sion of Fig. 2A,although this method does not nec-essarily provide correct photometric values. The brightest region in this case reaches -10 MR of H2 Lyman emission.
15. J. C. G6rardand V. Singh, J. Geophys. Res. 87, 4525 (1982).
16. H. J.Waite etal., ibid. 88, 6143(1983).
17. D.Rego, R.Prang6, J. C.Gerard, ibid. 99, 17075 (1994).
18. Theresults of the 1 W m-2 aurora excited by elec-trons with the Ulysses energyspectrum imply an H2+production rate exceeding 2 x 106 cm-3 with a peakproduction expected near the 1-ijbarpressure level and a peakH3+ density on the order of 108 cm-3. Thecorresponding heating rate also peaks at thispressure with a vertically integrated heat flux of over 450 ergs cm-2so-. The heating rate profile calculated from the model can be combined with recent cooling rate calculations (scaled so that the vertically integrated cooling rate equals the vertically integrated heating rate and peaks at a pressure level
of 1[ibar,justabove the methanehomopause, that
is,cool-to-space approximation). These values were used as inputs in a 1 Dthermal conduction equation (26) toestimate the auroral thermal profile. 19. L. M. Trafton, J. C. G6rard, G. Munhoven, J. H.
Waite, Astrophys. 1421, 816 (1994).
20. P. Drossart et al., J. Geophys. Res. 98, 18803 (1993).
21. J. T. Clarke et al.,Astrophys. J. Lett. 430, 73(1994). 1678
22. G. R. Gladstone and T. Skinner, NASA SP-494
(1989),pp.221-228.
23. M.Hammond andJ.Gosling, personal communica-tion.
24. M. K.Dougherty,D.J.Southwood,A.Balogh,E. J.
Smith,Planet.SpaceSci.41, 291 (1993); R.Prang6,
M. Dougherty,M. Dunlop,A.Balogh, V. Dols, Eos 74,43(1993).
25. V.Vasyliunas,Geophys.Res.Lett.21,401 (1994).
26. H. J.Waite,M. L.Chandler,R.V.Yelle, B. R.Sandel,
T.E.Cravens,J.Geophys. Res. 93, 14295(1988).
27. J. E. P.Connerney,inPlanetaryRadio Emission Ill, H. Rucker, M. L. Kaiser, S. J. Bauer, Eds. (Austrian Academy ofSciencesPress, Vienna, 1992), pp. 13-33.
28. J.C.G. acknowledges support from the Belgian
National Fund forScientific Research (FNRS) and
from the Belgian Federal Science Policy Office, Prime Minister'sServices.R.P. isgratefultoJ. Gos-ling,M.Hammond,K.Maeda,H.Oya,and P.Zarka forfruitful discussions and forhaving kindlyprovided
unpublisheddataontheJovian radio emission and
onthe solar wind. Sheacknowledges partial support
of thisstudy bygrantINSU-PNP/80374200.J.H.W.
acknowledgessupportfromSpace Telescope
Sci-enceInstitutegrantGO-4608.02-92A and National Aeronautics andSpaceAdministration(NASA)
Plan-etaryAtmospheregrantNAGW-3624. Thisstudyis basedonobservations with theNASA-ESA HST that
wereobtainedattheSpaceTelescopeScience
In-stitute,which isoperatedbyAURA forNASA under contractNAS5-26555.
11July1994;accepted27September1994
In
Situ
Determination
of
the NiAs
Phase of
FeO
at
High
Pressure
and
Temperature
Yingwei Fei and Ho-kwang Mao
In situ synchrotron x-raydiffraction measurements of FeO at high pressures and high temperaturesrevealed that thehigh-pressurephase ofFeOhas the NiAs structure (B8).
Thelatticeparametersof this NiAs phaseat96gigapascalsand800 kelvinarea=2.574(2)
angstromsandc= 5.172(4)angstroms(the number in parentheses is theerrorinthelast
digit). Metallic behaviorof the high-pressure phase is consistent with acovalently and metallicallybondedNiAs structure of FeO.TransitiontotheNiAs structure of FeO would enhanceoxygensolubilityin molten iron. This transition thus providesaphysiochemical basis for theincorporation ofoxygen into the Earth'score.
Ferrous oxide is a basic oxide component in the interiorofthe Earth. Its high-pres-sure and high-temperature behavior plays an important role in mantle composition and core formation models. Since shock wavestudies (1) revealedtheexistenceof a high-pressure phase of FeO at pressures above70GPa over adecade ago, the struc-ture of this phase has been the subject of speculation(1-4). Transition tothis phase has notbeenobservedinstaticcompression experiments in thediamond-anvil cell upto 120 GPa at room temperature (5). Addi-tional shock compression experiments (6) have confirmed that the density of FeO changes at 70 GPa. The discrepancy be-tween the static and shock compression experiments may be related to the experi-mental temperature difference in the two techniques. In this study, we report static high-pressure andhigh-temperature experi-ments onFeOthat resolve thisdiscrepancy.
Inaddition, insitusynchrotronx-ray
mea-surements reveal the structure of the high-pressure, high-temperature phase of FeO. We discuss implications of the high-pres-surestructure ofFeOforthe incorporation ofoxygen (in the form of FeO) into the Earth's core andfor the mineralogy ofthe lowermantle.
We havedevelopedanexternallyheated
Geophysical Laboratoryand Center forHigh-Pressure
Research,5251Broad BranchRoad, N.W.,Washington, DC20015, USA.
high-temperaturediamond-anvil cell(7, 8)
that iscapable ofachievingpressures
great-erthan 125 GPaattemperatures upto1100
Kin amildly reducing atmosphere (Arwith
1% H2). We conducted experiments on FeO by using this high-temperature cell combined with in situ synchrotron x-ray diffractionmeasurements.Atroom temper-ature,weconfirmed thetransitionfrom the cubic (B1) toa rhombohedral phase at 16 GPa under hydrostatic conditions (5) (Fig. 1). The transition occurs at much lower pressure (-8 GPa)whennopressure
medi-um is used in the experiments. High-tem-perature and high-pressure experiments, carried out inneonpressure medium, show that thetransition boundary hasapositive pressure-temperature slope with P = -5.0 + 0.070T (for pressure in gigapascals and temperature inkelvin).
At higher pressure, we carried out two
experimentsonFeOwithoutapressure me-dium. Thetwo experiments were identical in experimental configuration but different inthepressure andtemperature ranges.We observed no phase transition between 16 and 84 GPa at room temperature,
consis-tent with previous static compression
ex-periments (5). Uponheating, new diffrac-tion peaks started toappear at74 GPa and 900Kinthe firstexperimentandat90GPa and 600K inthe secondexperiment. Inthe first experiment, thediffraction peaks from
the new phase became more intense with
SCIENCE * VOL. 266 * 9DECEMBER 1994
m .Ml
on October 27, 2018
http://science.sciencemag.org/
Telescope
A Remarkable Auroral Event on Jupiter Observed in the Ultraviolet with the Hubble Space
J. C. Gérard, D. Grodent, V. Dols, R. Prangé, J. H. Waite, G. R. Gladstone, K. A. Franke, F. Paresce, A. Storrs and L. Ben Jaffel
DOI: 10.1126/science.266.5191.1675 (5191), 1675-1678.
266
Science
ARTICLE TOOLS http://science.sciencemag.org/content/266/5191/1675
REFERENCES
http://science.sciencemag.org/content/266/5191/1675#BIBL
This article cites 19 articles, 2 of which you can access for free
PERMISSIONS http://www.sciencemag.org/help/reprints-and-permissions
Terms of Service
Use of this article is subject to the
registered trademark of AAAS.
is a
Science
American Association for the Advancement of Science. No claim to original U.S. Government Works. The title
Science, 1200 New York Avenue NW, Washington, DC 20005. 2017 © The Authors, some rights reserved; exclusive licensee (print ISSN 0036-8075; online ISSN 1095-9203) is published by the American Association for the Advancement of
Science
on October 27, 2018
http://science.sciencemag.org/
